E182 ‘Revolutionising Medical Speech Recognition’ with T-Pro’s Meghan Dowling

Today’s guest is Meghan Dowling, Computational speech and language scientist at T-Pro. Meghan has gained recognition for her impactful work, including being a finalist in the 2024 AI Awards Women in AI Ambassador of the year category. Meghan speaks about her work within Automatic Speech Recognition (ASR) at T-Pro, an exciting company making strides in medical documentation.

Meghan’s journey into ASR began when she combined her passions for computer science and the Irish language, eventually leading her to a PhD in Irish machine translation. This path naturally progressed into applying her skills to medical contexts at T-Pro. Here, she helps to minimise the documentation burden on clinicians through advanced speech technologies, allowing more time for patient care.

For those unfamiliar with T-Pro, the company’s mission is to alleviate the heavy documentation workload that medical practitioners face. By integrating AI-driven solutions into daily practice, they aim to enhance productivity and ultimately improve patient care. As Meghan explains, over 50 percent of clinicians’ work involves mundane paperwork, which steals valuable time from their primary focus: patient interaction.

The accuracy of ASR technology is critical in the medical field. Any misrecognition could potentially jeopardise patient safety. Meghan highlights the importance of achieving the highest quality in ASR performance. Despite the challenges posed by specialised medical terminology and colloquialisms, T-Pro’s dedication to in-house systems ensures they maintain control and precision in recognising and converting speech to text.

One significant issue in AI today is the phenomenon of hallucinations and bias – AI making up data or misunderstanding accents, resulting in errors. T-Pro’s approach involves developing proprietary systems that avoid such pitfalls by using diverse and well-tagged datasets. As Meghan points out, the AI’s adaptability to various accents ensures everyone can use the technology without altering their natural voice, promoting inclusivity and improving user trust.

Throughout her career, Meghan has benefited from mentors who encouraged her to challenge norms and combat imposter syndrome. She draws inspiration from historical figures like Kay McNulty, one of the first digital computers’ pioneers, and actively promotes diversity at T-Pro. The company supports women in tech through diverse hiring practices and mentorship programs.

Looking forward, T-Pro is exploring the potential of a new ‘co-pilot’ project – an AI application working alongside doctors during consultations. This tool could generate detailed reports, letters, and prompts, assisting doctors in providing holistic patient care.

In conclusion, Meghan and the T-Pro team demonstrate how innovative applications of AI are transforming the healthcare industry. Meghan’s work not only improves efficiency for healthcare practitioners, but she is also a leader for diversity and inclusivity in the tech industry.